Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€) (@andrewcritchphd) 's Twitter Profile
Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€)

@andrewcritchphd

Let's make AI doctors!
Views my own;
CEO @ HealthcareAgents.com;
AI Researcher @ Berkeley;
If I block you it's like I'm moving to another convo at a party; nbd.

ID: 2457094022

linkhttps://acritch.com/ calendar_today21-04-2014 20:32:42

1,1K Tweet

4,4K Takipรงi

276 Takip Edilen

Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€) (@andrewcritchphd) 's Twitter Profile Photo

Twitter / X platform has finally learned after ~10 years that I like it when humans are kind to animals. For reference, it took TikTok about 40 minutes to learn this about me, and I think YouTube still doesn't know. Still, thanks X! Better late than never :)

Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€) (@andrewcritchphd) 's Twitter Profile Photo

The belief that ใ€Ša Python script can control what a computer doesใ€‹ is of the same type as the belief that ใ€Ša human mind can control what a human body doesใ€‹. You're welcome.

Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€) (@andrewcritchphd) 's Twitter Profile Photo

A lot of "AI alignment" work is too vague to be considered a coherent area of research. Check out Owain's results for refreshing counterexamples: actually-informative and clearly-exposited observations on LLM behaviors. Cheers, Owain!

Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€) (@andrewcritchphd) 's Twitter Profile Photo

4. "We" is vague and obfuscates who is expected to change. 5. Diffusion of responsibility is on track to accelerate AI industrial side-effects until humans, no longer competitive as producers or consumers of goods, are dead. 6. Someone or something could change *that*. Is it you?

Dawn Song (@dawnsongtweets) 's Twitter Profile Photo

1/ ๐Ÿ”ฅ AI agents are reaching a breakthrough moment in cybersecurity. In our latest work: ๐Ÿ”“ CyberGym: AI agents discovered 15 zero-days in major open-source projects ๐Ÿ’ฐ BountyBench: AI agents solved real-world bug bounty tasks worth tens of thousands of dollars ๐Ÿค–

1/ ๐Ÿ”ฅ AI agents are reaching a breakthrough moment in cybersecurity.
 In our latest work:

 ๐Ÿ”“ CyberGym: AI agents discovered 15 zero-days in major open-source projects

 ๐Ÿ’ฐ BountyBench: AI agents solved real-world bug bounty tasks worth tens of thousands of dollars
๐Ÿค–
Meredith Whittaker (@mer__edith) 's Twitter Profile Photo

'Meredith,' some guys ask, 'why won't you shove AI into Signal?' Because we love privacy, and we love you, and this shit is predictable and unacceptable. Use Signal โค๏ธ

'Meredith,' some guys ask, 'why won't you shove AI into Signal?'  

Because we love privacy, and we love you, and this shit is predictable and unacceptable. Use Signal โค๏ธ
Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€) (@andrewcritchphd) 's Twitter Profile Photo

I've said for years that the Hubble-visible region is itself probably a part or byproduct of a living system, simply because living systems grow and replicate, and thus on priors should be presumed prevalent. Glad to see the public ideation below from Andrej Karpathy and David .

Andrew Critch (๐Ÿค–๐Ÿฉบ๐Ÿš€) (@andrewcritchphd) 's Twitter Profile Photo

This is important. Environment really matters to health, fitness, and whether the meat you eat was tortured or lived a happy life. Bryan Johnson I'd love it if a "healthy living" hotel chain cropped up with your endorsement!